Reiner Pope
PersonHeard in 2 episodes across 1 show since Apr 2026
Reiner Pope is a technical expert who specializes in the mathematical foundations of training and serving Large Language Models (LLMs). He has appeared as a guest on the Dwarkesh Podcast to discuss memory bandwidth, compute efficiency, and the engineering behind AI systems.
Episodes
First heard
Guest appearances
Episodes per month
Public PodLume episodes featuring it, over the last year.
Show the data
| Month | Episodes |
|---|---|
| Nov 2025 | 0 |
| Dec 2025 | 0 |
| Jan 2026 | 0 |
| Feb 2026 | 0 |
| Mar 2026 | 0 |
| Apr 2026 | 1 |
| May 2026 | 1 |
| Jun 2026 | 0 |
| Jul 2026 | 0 |
| Aug 2026 | 0 |
| Sep 2026 | 0 |
| Oct 2026 | 0 |
What Reiner Pope has said on podcasts
68 statements
Matrix multiplication performs a multiply-accumulate at every step.
“multiply accumulate happens at every single step of a matrix multiply”
Open the episode · Reiner Pope – Chip design from the bottom upListen at 2:45
AI-chip accumulation generally requires higher precision than multiplication.
“the precision will almost always be higher in the accumulation step than in the multiplication step”
Open the episode · Reiner Pope – Chip design from the bottom upListen at 2:51
Accumulation causes errors to build quickly, requiring greater precision.
“when you accumulate errors accumulate quickly and so you need more precision here”
Open the episode · Reiner Pope – Chip design from the bottom upListen at 3:01
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Episodes
2 episodes featuring Reiner Pope, newest first
